Deduplication of Docker Image Registry

Containers are prominently being used today, as a mechanism for delivering application packages bundled with the requisite libraries, configuration information and everything needed to run and seamlessly deploy applications. Docker has evolved as the de-facto standard for building and sharing these containerized applications. Images, the bundled application templates, are executed with configuration information for creation of these application containers. These images stored in Docker registries, could be in DockerHub which is a public Docker registry or in private registries, have an increased need for storage with the increase in proliferation and adoption of containers. This increased need for scaling leads to the need for optimization approaches to manage and effectively use this storage. Analysis shows that there is an abundance of interrelationship, interdependence, common pieces and redundant components with these applications. This paper proposes a solution to reduce this redundancy using data deduplication and also proposed techniques to offset the potential increase in latencies for reconstructing images and thus a solution for optimal management of storage scaling.